Agent Control Protocol: Admission Control for Agent Actions

The Agent Control Protocol (ACP) addresses a critical gap in agentic security: emergent behavior. While tools like NeMo Guardrails or LlamaIndex provide per-request filtering, ACP focuses on the 'Ledger'—a stateful accumulation of risk signals over time. Despite the conceptual novelty of applying network-style admission control to LLM agents, the project currently lacks any market traction (0 stars, 1 fork). It exists primarily as a research artifact. Defensibility is extremely low because the core logic (anomaly accumulation and cooldown periods) is easily reproducible by any engineering team once the concept is understood. Frontier labs (OpenAI, Anthropic) and infrastructure providers (AWS Bedrock, LangChain) are the natural owners of this 'circuit breaker' logic. As agents become more autonomous, these platforms will inevitably ship native stateful monitoring to prevent reputation damage or liability from runaway agents, likely displacing this project within a 6-month window.